Flow control problems (Xon/X

[Charlie Dennett]

I support a 3/60 running 4.0.1.  The user of this system has a CAD CAM
package running on it.  One of the functions of this package is to send
the drawing to a plotter.  The plotter is attached to one of the two
serial ports on the 3/60 and is driven by the lpr(1) command.  From what I
can determine, the plot function stores in a spool directory some kind of
binary file.  At the beginning of this file are the commands for the -g
flag of the lpr(1) command.  This flag invokes a filter that runs a vendor
supplied program on the remainder of the spool file.  In effect, this
program converts the binary data into something the plotter can
understand.  In my case this is HPGL since the plotter is an HP7586B.

This whole setup works fine.  I've attached a breakout box to the RS-232
line.  From what I can tell it is using Xon/Xoff flow control.  I captured
the output of the filter to a file to look at the HPGL commands and
according to the manual for the plotter the first few commands set it up
for Xon/Xoff flow control.

Now for my problem.  I have a genealogy application I got from soc.roots.
One of the features of this application is to produce output in HPGL
format so the plotter can draw a family tree.  The first few lines are
identical to the ones I captured from the CAD CAM application mentioned
above.  However, if I take this and run it through lpr(1), flow control
does not work.  If the plotter does send back an Xoff, the 3/60 never sees
it.  The spool file is sent to the plotter in one continuous burst.
Needless to say, most of it end up in the bit bucket.

As another test, I took the HPGL data I captured from the CAD CAM
application and put it through lpr(1).  The same thing happened - no flow
control.  This leads me to believe that it is the vendor program that
translates from the binary form to HPGL that is doing the actual flow
control, not any of the lp* commands.  Is that true?  

What do I need to do to get Xon/Xoff to work?  I've included the latest
incantation of the /etc/printcap entry for the plotter.  I just looked at
the release manual for 4.0.3 and it appears that this might have been a
bug that was fixed in 4.0.3.  Is this true or am I doing something wrong.

The  printcap entry below is the latest version.  I've been playing with
the ms part but nothing has worked.

#plotter stuff for cimlinc


#
#***PLOTTER SERVER ENTRIES***
#
plotter:\
	:lp=/dev/plotter:\
	:rw:\
	:sh:\
	:sd=/var/spool/plotter:\
	:gf=/usr/bin/cimplot.sh:\
	:lf=/var/adm/plotter-errs:\
	:br#9600:\
	:ms=-parity,-cstopb,-clocal,cread,ixon:\
#	:if=/usr/lib/lpf:\
	:af=/var/adm/plotter-acct:

Thanks for any help.  BTW, if all this is in the manual, please tell me
where.  I've looked and can't seem to find it.

Charlie Dennett          | UUCP: ...!rutgers!rochester!kodak!cygnus!dennett
Information Services     | Internet:  dennett at cygnus.Kodak.COM
Eastman Kodak Company    | Phone: (716) 726-4480
Rochester, NY 14653-5219 | Minicomputer Tech. Support Group - HP3000/VAX/SUN